@inproceedings{78c04680bf304d98bd8e57341129f24b,
title = "GraphEBM: Energy-based graph construction for semi-supervised learning",
abstract = "With the rapid improvement of various techniques in graph-based semi-supervised learning, the call for higher-quality graphs becomes more intensive. However, such affinity graphs are not naturally existing in most semi-supervised learning tasks. In this paper, we propose a learning-based approach, GraphEBM, for the graph construction problem. GraphEBM is designed to address three main requirements in graph construction: 1) supporting dynamic update; 2) providing interpretable metrics; 3) tailoring to tasks. Specifically, in GraphEBM, we adopt a probabilistic view, Edge Probability Space, to model a graph construction process as constituted of events from the space. Our objective is thus to learn, by our Energy-Based Model (EBM), the latent sampling distribution. Experimental results show that our proposed GraphEBM outperforms the existing graph construction methods in improving the semi-supervised learning tasks on various datasets and it can learn global properties of a target graph only with direct local guidance.",
keywords = "Energy-based model, Graph construction, Graph semi-supervised learning, Probability space",
author = "Zhijie Chen and Hongtai Cao and Chang, {Kevin Chen Chuan}",
note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 20th IEEE International Conference on Data Mining, ICDM 2020 ; Conference date: 17-11-2020 Through 20-11-2020",
year = "2020",
month = nov,
doi = "10.1109/ICDM50108.2020.00015",
language = "English (US)",
series = "Proceedings - IEEE International Conference on Data Mining, ICDM",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "62--71",
editor = "Claudia Plant and Haixun Wang and Alfredo Cuzzocrea and Carlo Zaniolo and Xindong Wu",
booktitle = "Proceedings - 20th IEEE International Conference on Data Mining, ICDM 2020",
address = "United States",
}